GRUNDY, SCOTT M.; BLACKBURN, GEORGE; HIGGINS, MILLICENT; LAUER, RONALD; PERRI, MICHAEL G.; RYAN, DONNA
On February 4–6, 1999, the American College of Sports Medicine sponsored a scientific roundtable on the role of physical activity in the prevention and treatment of obesity and its comorbidities. The purpose of the conference was to provide an evidence-based review of the current state of knowledge on physical activity as a modality for coping with the “epidemic” of obesity occurring in the United States and other nations (1,5). Participants of the conference were requested to review existing literature and to classify available data according to accepted evidence-based categories. The categories employed are those outlined in the recent Clinical Guidelines on the Identification, Evaluation, and Treatment of Obesity in Adults reported by a task force of the Obesity Education Initiative (OEI) of the National Heart Lung and Blood Institute (NHLBI) (2) (Table 1).
A particular goal of this roundtable was to review randomized clinical trials (RCT) that provide evidence in Categories A and B. Unfortunately, there are a limited number of RCT of physical activity in obese populations. On the other hand, the field is rich in observational data and studies in human physiology and behavior (Category C). A sizable portion of the Category C evidence derives from large and replicated studies and provides a considerable base upon which recommendations can be made. Such recommendations indeed are set forth in the OEI report (2), the NIH Consensus Conference Statement on Physical Activity and Cardiovascular Health (December 18–20, 1995) (3), and the Report of the Surgeon General on Physical Activity and Health (4). The current roundtable aimed to extend these previous reports by examining prior and new publications in more depth and by categorizing evidence more precisely as to type and strength. The manuscripts upon which presentations and the current panel report are based are being published concurrently in Medicine and Science in Sports and Exercise. The glossary of key terms used in the current report are those derived from the Surgeon General’s report on physical activity and health (4) (Table 2).
Prevalence and trends in overweight and obesity.
The most recent National Health and Nutritional Examination Survey (1988–1994) (NHANES III) (1) revealed that 54.9% of American adults are overweight or obese (see Table 3 for OEI classification of overweight and obesity). Since 1960, overweight and obesity have increased across all ages, genders, and racial/ethnic groups. Prevalence in the obese category has increased by about 10%. Not only has the overall distribution of body mass index (BMI) shifted to higher levels, but the distribution has become even more skewed toward the high end. In the recent survey (1), the highest prevalence of obesity was found among non-Hispanic black women, Mexican-American women, Mexican-American men, and among less well-educated and low-income people.
The recent OEI report (2) emphasized that waist circumference is a good indicator of abdominal obesity, which at BMI levels ≤ 35 kg·m−2 is correlated more closely with the comorbidities of overweight than is BMI (see Table 4 for OEI classification of disease risk according to waist circumference). Cross-sectional epidemiological data are not available to define the prevalence of abdominal obesity in the population; neither are data available to correlate waist circumference with comorbidities in the whole population. Such data would add significantly to our understanding of the metabolic consequences of overweight.
Prevalence and trends in physical activity.
The 1996 Behavioral Risk Factor Survey (BRFSS) reveals a high prevalence of physical inactivity among American adults and high school students (see Table 5 for the classification of physical activity used in the Surgeon General’s report on physical activity). Adolescents are more active than adults; 64% of high school students report participating in vigorous (hard or very hard) activity for at least 20 min on three or more days per week. Boys engage in physical activity more than girls, and whites in general more than in blacks or Hispanics. Physical activity declines at higher grades in school, especially among girls. Among adults, only 28% of men and women achieve moderate or vigorous levels of physical activity. Further, 27% of men and 31% of women report no regular physical activity outside of work. Educational level is a factor in exercise pattern. Less than 20% of college graduates report being inactive, whereas nearly half the population with a high school education are inactive. Black and Hispanic men and women are less active than white men and women. In spite of long-term societal trends in activity patterns, these have not been assessed quantitatively; moreover, leisure-time activity apparently has not changed much over the recent decades. Future surveys might well examine the impact of changes in activity patterns at work and at home, and in active and passive leisure time activities (i.e., less active recreation during and after school and more television viewing, computer work, and playing video games).
The economic costs of obesity and inactivity.
The medical costs of obesity and physical inactivity can be estimated from the strengths of their associations with various diseases—coronary heart disease, diabetes, gallstone disease—and comorbid risk factors. The direct costs of a lack of physical activity, defined conservatively as absence of leisure-time physical activity, are approximately 24 billion dollars, or 2.4% of the U.S. health care expenditures. Direct costs for obesity defined as a BMI greater than 30 kg·m−2, in 1995 dollars, totals 70 billion dollars. The costs of inactivity are independent of obesity, and the costs of obesity are independent of those due to a lack of physical activity. Overall, the direct costs of inactivity and obesity are estimated to consume some 9.4% of national health care expenditures in the United States.
Determinants of overweight and obesity.
The high and increasing prevalence of overweight and obesity must be due in large part to environmental factors. Development of obesity requires that energy intake exceed energy expenditure; maintenance of obesity demands a higher energy input or a lower energy expenditure, or both, than needed for a healthier weight. Factors affecting both intake and expenditure of energy probably play a role in the causation and maintenance of obesity. Societal trends go against reducing energy intake as well as against increasing energy expenditure. A relatively high intake of energy is driven by a food supply that contains readily available, energy-dense foods, served in large portions. Energy expenditure is lowered by progressively lesser amounts of physical activity required at work and at leisure. Most energy expenditure is obligatory, determined by resting metabolic rate and the thermic response to food intake, but physical activity must not be discounted in the equation for total energy expenditure. Even in the absence of occupations or recreations that consume large amounts of energy, regular physical activity offers a means to lessen the severity of overweight and obesity in the population. A sustained increment in energy expenditure of 200 kcal·d−1 through increased physical activity would reduce body weight by about 5 kg over a period of 6 months to 1 yr, assuming no increase in food consumption. Unfortunately, any weight loss achieved by moderate physical activity can be easily reversed by small compensatory increases in food intake.
Measurements of body habitus and energy parameters.
If the role of physical activity in the causation or treatment of obesity is to be placed on a quantitative basis, accurate measurements must be made of body composition, energy intake and expenditure, and levels of physical activity. Methods for measurements in each of these areas are improving but still have significant limitations. Significant strides have been made in measurements of body composition—total body fat, body fat distribution, lean body mass, and muscle mass. Several of these techniques are increasingly being used in epidemiological studies, but the most sophisticated measurements can be costly. Despite the advantage of being safe enough for use in children and pregnancy, the latter are too laborious and costly for use in large populations. Methods for measurement of nutrient and energy intake leave much to be desired. Questionnaires and diaries of food intake are plagued by under reporting. Precise methods for estimating energy intake, as would be required to define differences in intake responsible for weight gain, are not available.
The most objective and accurate method for assessing the level of physical activity and energy expenditure of activity is average daily metabolic rate, determined by the doubly labeled water, minus the basal metabolic rate. This method, however, is limited to studies in small numbers of subjects. More applicable to population studies are motion sensors, specifically, accelerometers. These instruments reliably assess patterns of physical activity. Accelerometers offer the advantage of measuring motion from nonexercise (i.e., lifestyle) activity as well as exercise activity.
Another way to estimate “integrated” physical activity is to measure cardiorespiratory fitness. This measure is represented by maximal oxygen consumption determined under exercise conditions. This measurement carries the advantage of being quantitative and available for epidemiological studies. It is not, however, a direct measure of physical activity but only a reflection of it. Moreover, cardiorespiratory fitness is largely an indication of recent intensive exercise; it may not be a useful tool to assess the effects of moderate exercise, and it is also influenced by inherited characteristics.
PHYSICAL ACTIVITY IN THE ETIOLOGY AND TREATMENT OF OBESITY
Sedentary life habits and inactivity in the etiology of overweight and obesity.
No RCT are available that address whether sedentary life habits and inactivity contribute to the development of obesity in populations. Available data therefore are restricted to observational studies (Evidence Category C). Studies of ecological trends in populations provide suggestive evidence that declining amounts of physical activity have contributed importantly to a rising prevalence of overweight and obesity. Some studies have reported secular decreases in energy intake concurrently with increases in body weight, both in children and adults; these decreases infer a corresponding even greater decrease in energy expenditure, which would be a requirement for weight gain.
Two cross-sectional studies, each on about 700 children, noted lower levels of activity being correlated with higher levels of body fatness. In these studies, confounding variables could not be fully excluded. Children of parents with less formal education and lesser incomes had greater body fatness that was associated with more watching of television and other sedentary activities. In another cross-sectional study, Minnesota researchers used questionnaires to obtain information about food intake and exercise frequency on large samples of women and men. Body weight correlated inversely with high intensity exercise in men and with both high intensity exercise and walking in women. At the same time, high consumption of certain foods and drinks—alcohol, dairy products, and meat—were positively associated with body weight. This study supported a dual etiology for increased body weight.
A review of eight prospective studies revealed some link between low levels of physical activity and risk of developing obesity. In one study in U.S. male health professionals, those reporting vigorous exercise had lower body weights at follow-up. There was a direct association between television viewing habits and body weight. Relationships were more pronounced in middle-aged men than elderly men. In the Women’s Gothenburg study, low levels of leisure-time physical activity appeared to be a risk factor for weight gain in women consuming high-fat diets. Taken together, the various observational studies support an inverse relationship between the level of physical activity and body fatness. This support resides in Evidence Category C, and it can be called only moderately strong—the data suffer from a lack of precision in measurements of energy intake and total physical activity.
Prevention of weight gain through physical activity.
RCT are not available to address whether physical activity can prevent weight gain in the general population. Data are restricted to observational studies (Evidence Category C). The age-related pattern of body weight demonstrates an increasing prevalence of weight gain through the sixth decade. Several large-scale longitudinal studies have assessed the role of habitual physical activity in preventing the progression of weight gain from acceptable weight through overweight to obesity, along with its attendant comorbidities. These studies suggest that physical activity and fitness attenuate weight gain but do not necessarily prevent weight gain or promote weight loss. Fitness (and maintaining fitness) further appears to protect against sizable weight gain. Although body weight increases with advancing age, even among the most fit and highly active, this increase may be attenuated through physical activity. Avoiding a sizable weight gain with aging through physical activity nevertheless may mitigate comorbidities of obesity. A considerable body of data exists that physical activity can attenuate weight gain with aging with attendant benefits in reduction of comorbidities (Evidence Category C).
Treatment of overweight and obesity in adults.
Ten RCT have addressed whether exercise interventions alone will produce weight loss. The majority of these studies show that physical activity alone in the form of aerobic exercise produces some weight loss. The effect of exercise alone on weight loss nonetheless appears to be modest—generally about 1–2 kg over the duration of study. Evidence from this group of RCT was not considered robust enough for Evidence Category A. Several studies were relatively small, and on the whole, observed weight changes were modest, although consistent. The panel therefore assigned the summed results to Evidence Category B.
Another 13 RCT examined whether the combination of reduced-energy diets and increased physical activity produce a greater weight loss than does a low-energy diet alone. Most studies favored diet + exercise regimens over diet alone, but the difference was statistically significant in only a minority of studies. In one of the latter, diet + exercise proved better in men, but not in women. The panel concluded that the overall lack of statistical significance was probably due to the short time-frame of the exercise programs, small sample sizes with inadequate statistical power, and difficulty with adherence to exercise. These limitations in design placed the summed results of this group of RCT in Evidence Category B.
A few of the prior RCT were extended to examine whether a regimen of diet + exercise produces better maintenance of weight loss than does diet alone. Efficacy data from RCT in free-living adults are limited by small sample sizes and by poor long-term adherence to the recommended exercise and diets. Of six RCT that bear on the question, two showed significant long-term effects favoring diet + exercise over diet alone; the others were inconclusive. The relatively strong trends of the two positive studies lead the panel to conclude that diet + exercise provides benefit over diet alone in maintaining weight loss (Evidence Category B). This conclusion is bolstered by correlational analyses as well as studies of successful weight losers, which consistently show that physical activity is strongly associated with better long-term maintenance of weight loss (Evidence Category C).
Treatment of obesity in children and adolescents.
Obesity acquired in childhood appears to be predictive of adult obesity; therefore, prevention of obesity in childhood and adolescence is needed. Six RCT have tested whether low energy diets + exercise produce better weight loss than diet alone in children. The RCT were relatively small and were not powered to provide a definitive result. Taken together the RCT demonstrated that exercise incrementally adds to low-energy diets for short-term changes in percent overweight or percent body fat (Evidence Category B). The evidence further suggested to the panel that whereas exercise in overweight children and adolescents has limited effects on percentage overweight, it still improves body composition. The resulting fitness favorably modifies cardiovascular risk factors (Evidence Category D).
Physical activity for prevention and treatment of weight gain in pregnancy.
RCT were not uncovered that tested whether increased physical activity prevents or attenuates excessive weight gain during pregnancy. Few observational studies of exercise in pregnancy have examined maternal weight gain as an outcome. One prospective study compared women who chose to remain physically active during their pregnancies with women who chose to terminate their exercise habits. During the third trimester, exercising women showed reduced rates of weight gain and significantly smaller increases in fat accumulation (as measured by the sum of five skinfolds) compared with nonexercising women. Similarly, according to an observational study, women who had lower levels of leisure-time physical activity retained more of their pregnancy-associated weight gain than those who exercised. These observational studies were considered to be suggestive of benefit of exercise for weight control both during pregnancy and afterward and were assigned an Evidence Category C. The panel concluded that the development of physical activity interventions in pregnant women represents a potentially fruitful avenue for the prevention of overweight and obesity in women (Evidence Category D).
Physical activity and changes in body weight and fat distribution at menopause.
Weight gain during and after menopause contributes to overweight and obesity in older women. Three RCT have compared an exercise regimen with no intervention and another three RCT have compared diet + exercise with diet alone on body weight and fat distribution in postmenopausal women. The results of these RCT do not allow a firm conclusion as to whether physical activity will prevent or limit a gain in total body fat or abdominal fat after the menopause, or whether exercise provides incremental benefit over diet alone in an obesity treatment program for postmenopausal women. The available RCT were deemed to be insufficiently powered to adequately address this question. On the other hand, observational studies, both cross-sectional and longitudinal, suggest that postmenopausal women with high levels of physical activity have lower body fat and less abdominal fat during the menopause than do those with low levels of physical activity; exercisers moreover are less likely to gain in total body fat and abdominal fat (Evidence Category C).
Several interventional studies have examined whether exercise in the older population can slow down (or reverse) progressive depletion of muscle and accumulation of fat that occurs with advancing age. The panel considered intervention studies of at least 2 months’ duration and that included exercise, but not diet. Patients ranged in age from 55 to 86 yr and underwent either resistance training or aerobic training. Randomization occurred in about one-fourth of studies with aerobic training and in about half with resistance training. Aerobic training reportedly reduced fat mass without changing fat-free mass; the magnitude of loss of fat mass depended on duration of the trial. Resistance training reduced fat mass and increased fat-free mass, unrelated to the duration of the exercise program. The panel concluded that resistance training reduces fat mass and increases muscle mass in persons over age 55, whereas aerobic training reduces fat mass but does not increase muscle mass (Evidence Category B).
Physical activity and abdominal obesity.
Five RCT examined whether increased physical activity alone, without diet-induced weight loss, will reduce body weight and abdominal girth. In four of five RCT, addition of exercise without dietary change failed to reduce body weight, body fat, or abdominal obesity. In only one RCT did exercise alone produce significant weight loss and decrease in abdominal circumference. In nine nonrandomized trials, the same question produced similar trends: exercise alone produced little or no change in body weight, body fat, or abdominal circumference. However, these 14 studies were not designed to determine whether there was a preferential reduction in abdominal fat. When visceral and subcutaneous abdominal fat are measured outcomes, exercise has been demonstrated in one randomized trial and four nonrandomized trials to result in reductions in both compartments. The panel concluded that increased physical activity alone without weight loss is not associated with reductions in abdominal girth. Physical activity with or without weight loss is associated with reductions in visceral and abdominal subcutaneous tissue (Evidence Category C).
OVERWEIGHT, OBESITY, AND COMORBID CONDITIONS
Obesity-associated mortality rates.
Prospective studies show the relationship between BMI and total mortality to be J-shaped, with minimal mortality occurring in the BMI range of 18.5–24 kg·m−2. Smoking along with overt and subclinical illnesses partially confounds the increased mortality at the lowest BMI. By excluding deaths in the first 5 yr and by confining the analysis to healthy nonsmokers, the studies reveal a linear relationship with mortality increasing progressively from lowest to highest levels of BMI. Multiple factors influence mortality rates and may confound the relationship between physical activity and reductions in premature mortality (Evidence Category C). This category of evidence is strengthened by a study in twins in which leisure-time activity correlated positively with lower mortality, after controlling for genetic and familial factors. These various studies still may not adequately control for life habits, risk factors, and socioeconomic status, all of which are correlated with BMI and mortality rates. Causes of death and life expectancy moreover have changed over time; birth cohorts have been exposed to different environmental factors that further complicate the interpretation of data. These confounding factors however almost certainly do not negate the causal relationship between BMI and mortality, which is linear and strong (Evidence Category C).
Strong epidemiological and metabolic data demonstrate that obesity contributes to a number of medical conditions: insulin resistance, glucose intolerance, diabetes mellitus, hypertension, dyslipidemia, sleep apnea, arthritis, hyperuricemia, gallbladder disease, and certain types of cancer. This evidence, which belongs to Category C but is overwhelmingly strong, has been reviewed in detailed in the recent NHLBI OEI report (2). The contribution of obesity to several cardiovascular endpoints—coronary heart disease, heart failure, cardiac arrhythmia, and stroke—appears to be independent of the influence of obesity on known risk factors. Many epidemiological and metabolic studies confirm an influence of obesity on risk factors related to insulin resistance and upon the development of Type II diabetes. This influence is augmented when the excess body fat is distributed predominantly to the abdomen and trunk (Evidence Category C). In spite of the strong associations between obesity and its comorbidities, much remains to be learned about underlying mechanisms and quantitative aspects of these associations.
Exercise and insulin resistance, impaired glucose tolerance, and diabetes.
Intervention-based RCT consistently show that increased physical activity improves insulin action, and thus reduces insulin resistance, in obese subjects (Evidence Category A). Epidemiological data suggest that exercise alone and exercise combined with weight reduction retard the transition from impaired glucose tolerance to Type II diabetes (Evidence Category C). For treatment of hyperglycemia in patients with diabetes, intervention studies on the value of exercise are mixed. Although exercise enhances insulin sensitivity, many patients with diabetes cannot engage in high-intensity exercise. Further, when defects in insulin secretion are severe, the treatment of hyperglycemia shifts toward providing more insulin to overcome insulin deficiency. Of note, any benefits from exercise bouts are short lived; therefore, if exercise is to be an effective adjunct in the long-term control of hyperglycemia, it must be regular and sustained.
Observational studies reveal that the major lipoprotein abnormality accompanying obesity is atherogenic dyslipidemia (elevated triglycerides, increased small low density lipoprotein (LDL), and low high density lipoprotein (HDL) levels). An elevated serum LDL cholesterol, when it occurs in overweight persons, appears to be related more to high intakes of cholesterol and cholesterol-raising fatty acids (saturated and trans fatty acids) than to overweight per se. Beyond observational studies, 11 RCT of 5–12 months’ duration and of large enough sample size to detect significant changes in HDL-C (and possibly other lipoproteins) have been carried out. These studies indicated that in the absence of weight loss exceeding at least 2.5 kg, an increased physical activity rarely raises HDL cholesterol or lowers triglycerides in overweight women (pre- or post-menopausal) or men. However, if physical activity is of a volume that results in at least 4.5-kg weight loss, HDL cholesterol will be raised, and triglycerides lowered, in men and postmenopausal women. Further, the addition of physical activity to a low-energy, low-fat diet will reverse the HDL-lowering effect of the low-fat diet in overweight men and women. Finally, in men and postmenopausal women with atherogenic dyslipidemia, the addition of physical activity to a modestly weight reducing, low-fat diet significantly enhances the LDL-lowering effect of the diet. The panel considered the RCT evidence for these conclusions to be strong (Evidence Category A).
A review of 44 RCT, involving 68 study groups and including 2674 participants, revealed that dynamic aerobic training reduces blood pressure by −3.4/−2.4 mm Hg (P < 0.001), after controlling for weight loss and dietary change and after weighting for the number of participants in each group. Blood pressure was reduced significantly more in groups of hypertension patients (−7.4/−5.8 mm Hg) than in normotensive study groups (−2.6/−1.8 mm Hg). Decreases in blood pressure were independent of weight loss and were not related to initial body mass index. Finally, eight RCT indicated that exercise alone is less effective for reducing blood pressure than diet alone (−3.5/−2.9 vs –5.4/3.7 mm Hg, respectively, P < 0.02); moreover, adding exercise to a low-energy diet seemingly does not reduce blood pressure significantly more than does diet alone. The panel concluded that RCT provide strong evidence that physical activity can reduce blood pressure in both lean and obese subjects. Furthermore, blood pressure reduction is independent of weight loss or changes in body composition, although the response does not go beyond the blood pressure reduction obtained by diet alone (Evidence Category A).
Cross-sectional surveys further suggest an inverse relationship between habitual physical activity or measured physical fitness and blood pressure. Three prospective studies agree that physically active subjects have both lower blood pressure and a reduced incidence of hypertension; these beneficial effects appears to be independent of a number of confounding factors including BMI or body fat. In one study, Harvard alumni who did not engage in vigorous play at sports in postcollege years displayed a greater risk for developing hypertension than did those who did play; this difference was observed only in those who were overweight at baseline. The panel agreed that observational studies support the concept that physical activity reduces the risk for developing hypertension, independent on its effects on body size or body composition (Evidence Category C).
A prothrombic state is one component of the metabolic syndrome that has been reported to occur in obese patients. The effects of physical activity on the coagulation system are complex and not well understood. Physical activity has an acute effect of promoting thrombosis and activating fibrinolysis. This effect might explain the acute coronary events that occasionally occur during intense exercise. Limited data further suggest an effect of physical activity on coagulation factors over a longer term. For example, cross-sectional studies suggest an inverse relationship between levels of physical activity and plasma fibrinogen levels. Still, intervention RCT and several noncontrolled studies remain contradictory, whether the endpoint is fibrinogen, plasminogen activator inhibitor, tissue plasminogen activator, or other coagulation factors. Taken as a whole, observational studies are modestly suggestive that regular physical activity is antithrombogenic (Evidence Category C). Limited RCT data further suggest that moderate-intensity physical activity carried out on a regular basis reduces platelet aggregation (Evidence Category B).
Many observational studies have found an association between obesity and some forms of cancer, notably, cancers of the colon, breast, endometrium, and possibly the prostate gland. Whether obesity plays a causative role in the development of these cancers is uncertain. Epidemiological studies also find a strong inverse relationship between physical activity and colon cancer. The strength and consistency of this association is so strong that a causal connection seems likely (Evidence Category C).
Prospective studies of physical activity and morbidity/mortality in overweight or obese individuals.
Among over 700 articles related to this topic, 25 were selected as providing data on the relation of physical activity to morbidity and mortality in strata of overweight and obese individuals. One question addressed was whether higher levels of physical activity or cardiovascular fitness reduces morbidity, mortality, and disability in overweight or obese persons. The panel discerned that active and fit individuals appear to have lower rates of disease and death than inactive persons who are not physically fit. This worsening gradient was present in various strata of body habitus. A second question was whether overweight and obese persons who are physically active have a lower risk of morbidity, mortality, and disability than normal weight individuals who are sedentary. The data indicated that being physically active and fit reduces obesity-related chronic diseases and decreases risk for early death and that active and fit persons who are overweight or obese actually have lower morbidity and mortality risk than normal weight persons who are sedentary. The evidence related to these questions was considered to be moderately strong and was classified under Evidence Category C.
The prevalence of overweight and obesity is alarmingly high and is increasing in the United States and many other nations (5). These conditions impose a heavy burden of morbidity and premature mortality as well as financial costs. Overweight and obesity are due to a relatively high intake of energy that is not matched with appropriate energy expenditure. One component of energy expenditure comes from exercise and other forms of physical activity. That most people in our society live sedentary lives is obvious to the casual observer and is a conclusion supported by observational studies. On the other hand, the solution to the problems of overweight and obesity cannot come exclusively from the institution of regimens of moderate exercise in the general population. Increments in energy expenditure brought about by moderate exercise are insufficient to completely forestall weight gain with advancing age or to reverse higher levels of overweight and obesity. Further, current social structures are not conducive to marked increments in energy expenditure through increased physical activity. Thus, curtailment of energy intake and efforts to increase physical activity will both be necessary to cope with the “epidemic” of overweight and obesity.
Nonetheless, a large body of evidence has been acquired to support the need to include increased physical activity in a program of weight control. Some of this evidence comes from RCT, although there have been too few such studies in the exercise field. Many observational studies, however, lend strong support to a place for regular physical activity in weight control. They suggest that a decline in physical activities in both occupational and leisure-time settings contributes to increasing overweight and obesity and that a restructuring of social conditions to allow and encourage physical activity would help to reverse current trends. Several RCT further indicate that combining a physical activity regimen with low-energy diets facilitates weight loss and maintenance of lower body weights. Finally, increased physical activity appears to have an independent beneficial effect on several comorbidities of obesity, notably insulin resistance, hyperglycemia, dyslipidemia, and, possibly, a prothrombic state. In addition, an active and fit way of life appears to attenuate morbidity and mortality risk in overweight and obese individuals.
Continued monitoring of patterns and trends in physical activity is needed for representative samples of the population. Improved methods of measurement appropriate for population-based observational studies are also required. Increasingly evidence-based recommendations require RCT to bolster the results of observational studies. Various epidemiological studies—cross-sectional, case-control, and prospective—may uncover new relationships and correlations. These can be extended by in-depth investigations of human metabolism and physiology. Nonetheless, in some crucial areas, causal relationships and clinical benefit of therapeutic modalities can be established only through RCT. The fields of physical activity and exercise physiology are rich in observational studies and research in human metabolism and physiology. On the other hand, they are lacking in RCT, although in recent years, an impressive number of RCT are being reported. It nonetheless can be said that one of the major needs in the field of physical activity is for more and better-designed RCT of physical activity in overweight and obese groups.
If results of RCT on the benefits of physical activity are to be convincing to the medical community, they must be based on sound methodology. They must have sufficient sample size, duration, and statistical power to yield a definitive result. The volume of exercise in the “treatment arm” of the study must be high enough to significantly affect metabolism and physiology. Adherence to the exercise intervention must be documented. The endpoints of the study should be clearly defined and be expressed in quantitative terms. Among endpoints, effects of intervention on magnitude of weight loss versus maintenance of weight loss should be differentiated. The experimental groups should contain both men and women; many previous trials appear to have been lacking in women as subjects. New RCT should be designed to test both efficacy (dose-response) and effectiveness (based on intention to treat) of the intervention. Development of appropriate control groups is mandatory; for example, consideration might be given to comparing different intensities or types of physical activity, rather than a simple comparison of exercise versus sedentary groups.
If RCT designed to evaluate the impact of physical activity on body weight are to be meaningful, quantitative methodology will be required. A great need is for the ability to measure total, 24-h physical activity. Current methodology is inadequate, although accelerometer technology is promising. The use of “physical fitness” as a surrogate for integrated physical activity needs more validation; of course, physical fitness may be an independent parameter related to health outcomes. The field of physical activity shares with the nutrition field the need for accurate measurements of energy consumption. The doubly labeled water technique is the “gold standard” for estimating energy (and intake in a weight-stable condition) but is impractical for use in epidemiological studies. Newer quantitative methodology is needed. Thus, if RCT are to take their rightful place in the physical-activity field, sound methodology for quantitative measurement is a necessary underpinning. Improvements in methodology will allow a more accurate assessment of the potential benefits of physical activity in the treatment of obesity and its comorbidities.
© 1999 Lippincott Williams & Wilkins, Inc.